Big Tech S3E14: Kate Crawford On the Toll AI is Taking on Humans and the Planet

May 27, 2021

Listen to this week’s new episode of Big Tech, where Taylor Owen speaks to Kate Crawford, research professor of communication and science and technology studies at USC Annenberg and a senior principal researcher at Microsoft Research Lab – NYC, about how AI development has been met with questions about its many impacts on our world and in light of this, whether we should consider if it’s worth developing it at all.

Artificial intelligence (AI) is hailed as a great technological leap forward, one that will enable immense efficiencies and better decision making across countless sectors. But AI has also been met with serious questions about who is using it and how, about the biases baked in and the ethics surrounding new applications. And an ever-growing concern about AI is the environmental toll it takes on our planet. Do the benefits of AI innovations outweigh all these concerns? Is it even worth it to develop AI at all?

In this episode of Big Tech, Taylor Owen speaks with Kate Crawford, research professor of communication and science and technology studies at USC Annenberg, a senior principal researcher at Microsoft Research Lab – NYC and author of the new book Atlas of AI. Crawford studies the social and political implications of AI.

Crawford’s work gets to the core of AI, looking at it as an extractive industry. AI extracts data, but it also extracts labour, time and untold quantities of natural resources to build and power the vast banks of computers that keep it running. Crawford argues that much of the work in AI is not, in fact, built in some “algorithmic nowhere space” on pure data, objective and neutral, but instead grounded on the ghost work and human labour that trains these systems. The industry mythologizes the internal workings, “our deeply abstract mathematical and immaterial understanding of AI,” as a way to avoid scrutiny and oversight. As Crawford explains, rather than try to govern AI as a whole, we need to take a broader approach addressing the many extractive aspects of AI to effectively tackle this problem and its wider planetary costs.